Visual Browser

About the application

Visual Browser is a Java application that can visualise the data in RDF scheme. The main principle of the visualisation is that:

Visual Browser uses the Jena framework to obtain the data, since the RDF scheme can be saved in different forms (a single XML file or a relational database).

The visualisation engine is derived from TouchGraph LLC.

The graph is animated to create an impression of fixed points (nodes) and elastic connections (edges). The main advantage of this method is that one node can be connected to a large number of other nodes without solving the problem of nodes obscuring their neighbours.

The user interface allows the user to expand and hide nodes, switch the view of edges and display hints. At any moment only a small part of the network is focused. Therefore the risk of getting the user “lost in hyperspace” is minimized.

Animated expansions let the user know the actual changes in the graph. This feature lets the user know what happens while changing from one view to another.

The whole visualisation engine corresponds to the navigation approach of WebOFDAV.

Two-level visualisation

Visual Browser has two levels of visualisation: the data and the so-called perspective of view. While the RDF scheme is very general, the perspective is bound to the particular data. It is a simple XML file, which contains information about how to display different nodes, edges and hints.

Examples of visualisation

EuroWordNet visualisation

EuroWordNet has two kinds of objects: synsets and literals. There are about 20 types of relations between synsets and a relation between a word and a synset.

Different perspectives allow to look EuroWordNet data from different points: either as a thesaurus, or as a dictionary

Visualisation of WordNet data can show many relations between concepts and words in one snapshot

Visualisation of WordNet data can show many relations between concepts and words in one snapshot

EuroWordNet as a dictionary. The clue is generated automatically depending on the perspective

EuroWordNet as a dictionary. The clue is generated automatically depending on the perspective

In complex data, at least some petterns can be observed. This is a visualisation of the word play, some of its meanings and some related concepts.

In complex data, at least some petterns can be observed. This is a visualisation of the word play, some of its meanings and some related concepts.

Library catalogue visualisation

In the library catalogue, there are objects like author, article, journal, keyword etc. Relations among all these objects are very natural.

Example of library catalogue visualisation with appearance of the hint

Example of library catalogue visualisation with appearance of the hint

Another example of the library catalogue: nodes representing authors have different color (orange), while articles are represented with blue points. Keywords are gray.

Another example of the library catalogue: nodes representing authors have different color (orange), while articles are represented with blue points. Keywords are gray. The library visualisation connects similar topics through authorship and mostly through the keywords.

Java source code visualisation

Visualisation of the source code

Programs for analysing compiled Java code help to retrieve many information from a compiled bytecode – packages, classes, methods, variables and the dependencies. Dependency finder is one of such tools. It can output the dependency graph. With a simple XSL transformation, a RDF from the dependency graph can be created. With the appropriate RDF and the perspective http://nlp.fi.muni.cz/projekty/visualbrowser/dependencies.xml one can browse dependency graph of an arbitrary Java program.

Run Visual Browser

Installation

You can download the whole package. After unzipping, run java -jar visual.jar.

Another way is to run Visual Browser through the Java Web Start. If it does not work or you don't know, what Java Web Start is, try to download and install Java Web Start from Sun. Note that Java Runtime Environment 1.5.0 beta has a bug and if you fail to run the application with 1.5.0 beta, try to upgrade your JRE.

Usage and sample data

You can to load the data – from a file, an URL or a database. Supported databases are MySQL and Postgre SQL. The database scheme must respect the Jena scheme. A small example data is on http://nlp.fi.muni.cz/projekty/visualbrowser/library.rdf (data from the library catalogue).

To load a perspective, click either on Open Perspective or Open Perspective from URL. The perspective appropriate to the library data can be found on http://nlp.fi.muni.cz/projekty/visualbrowser/library.xml.

Java Tutorial

Another ontology can be found at http://cohse.semanticweb.org/ontologies/java. The appropriate perspective can be found at http://nlp.fi.muni.cz/projekty/visualbrowser/java.xml. This ontology illustrates the Java Language Tutorial from Sun Microsystems page.

EuroWordNet

A small snippet of czech and english EuroWordNet can be opened from http://nlp.fi.muni.cz/projekty/visualbrowser/ewn.rdf. There are 2 perspectives: english thesaurus at http://nlp.fi.muni.cz/projekty/visualbrowser/thesaurus.xml and czech-english dictionary at http://nlp.fi.muni.cz/projekty/visualbrowser/dictionary.xml. The data is very small, there are words like man, woman, girl, lady etc. English words are in the namespace http://nlp.fi.muni.cz/wn/en, while czech words are in the namespace http://nlp.fi.muni.cz/wn/cz.

Citeseer

Citeseer is a well-known database of bibliographical information. A small part of citeseer data is available from http://nlp.fi.muni.cz/projekty/visualbrowser/citeseer.rdf. The appropriate perspective is available from http://nlp.fi.muni.cz/projekty/visualbrowser/citeseer.xml. Concepts (articles and authors) are connected by (co-)authorship and by citation. We assume that one author has a small range of themes he writes about and we assume that articles cited in another article are valuable in that theme.

Credits, references

to be updated

Valid XHTML 1.0!